Kidney and Urinary Tract Polyomavirus Infection and Distribution: Molecular Biology Investigation of 10 Consecutive Autopsies

BK virus (BKV) and JC virus (JCV) are human polyomaviruses (PVs), which infect most people throughout the world, as shown by the presence of antibodies against viral proteins in about 80% of the human population.1 

Although a large number of studies have been carried out, the mode of transmission, mechanisms of PV infection, and the distinct organs and cell types targeted by the viruses in different phases of infection have not yet been completely clarified. This is particularly true for the target organs, where the viruses persist indefinitely in a latent state.2 Indeed, searches for the sites of viral latency have led to discordant results: DNA sequences of JCV, BKV, or both have been detected in almost all the organs (kidney, brain, lung, bone, spleen, placenta) by a few authors,3–5 but these results have not been confirmed by other researchers, who identified the kidney as the selective site of JCV and BKV latency and the brain as a site of JCV latency.6–8 

Immune impairments are involved in the reactivation of PVs,9 leading to a wide spectrum of clinicopathologic consequences: in particular, progressive multifocal leukoencephalopathy (PML), a JCV-induced disease, has been frequently found in acquired immunodeficiency syndrome (AIDS) patients,10 and some diseases involving kidney and urinary tract, such as polyomavirus nephropathy (PVN),11 hemorrhagic cystitis, and ureteral stenosis, have been reported in kidney transplant patients12 and only rarely in AIDS patients.13 

In the case of the latter group of diseases, BKV has been considered the only etiologic agent of renal damage14 and ureteral stenosis,15 whereas hemorrhagic cystitis has been attributed from time to time to both JCV and BKV.16,17 This could suggest that not only various organs, but also different segments of the same apparatus or cell types, may be selectively infected by different types of PVs, and that the reactivation induced by immune impairment can cause organ-specific damage.

In an attempt to clarify this last hypothesis, kidney tissue and samples from various segments of the urinary tract (renal pelvis, ureter, and urinary bladder) obtained during 10 consecutive autopsies were investigated by means of molecular biology in order to detect the presence and distribution of the PV genome; moreover, in situ hybridization and immunohistochemistry were also carried out to define the viral status of the infected tissues.

The study involved complete autopsies of 10 unselected subjects consecutively performed between January and March 2003; the demographic and main pathologic data are shown in Table 1. It is worth stressing that, as shown by their clinical histories, all the subjects were immunocompetent except for patient 5, who was affected by AIDS.

Seventeen samples for each case (N = 170) were taken from macroscopically undamaged areas of the kidney and urinary tract as follows: 2 cortical and 2 medullary samples of each kidney (n = 80 samples), 1 sample of each renal pelvis (n = 20 samples), 2 samples of each ureter (n = 40 samples), and 3 samples of the urinary bladder (n = 30 samples). All of the samples were fixed in 10% buffered formalin, embedded in paraffin, and routinely processed for histology. Four-micrometer-thick sections were stained with hematoxylin-eosin and examined by means of light microscopy in order to evaluate the integrity of the tissue before proceeding to molecular analysis, to identify possible pathologic changes, and in particular to search for the presence of morphologic equivalents of cellular PV infection (intranuclear viral inclusions of various types, as described by Nickeleit et al18).

The samples were stained with immunoperoxidase in order to identify the presence and (latent or replicative) status of PVs using a monoclonal antibody against simian virus 40 (SV40) T antigen (clone pAb416, Oncogene Science, Boston, Mass; dilution 1: 60) and a polyclonal antibody against capsid proteins SV40 (Lee Biomolecular Research Labs, San Diego, Calif; dilution 1:20 000), both of which cross-react with human BKV and JCV.19 The reactions were detected by means of the streptavidin-biotin method and were revealed using diaminobenzidine as a chromogen. Brain sections with histologically proven PML from case 5 were used as a positive control.

In situ hybridization was performed to localize the nucleic acid sequences of BKV and JCV at the subcellular level using commercially available biotinylated DNA probes (Enzo Diagnostics, New York, NY). The reactions were detected by means of the streptavidin-biotin method and were revealed using diaminobenzidine as a chromogen. Brain sections of case 5 with PML and kidney sections of histologically proven BKV nephropathy were used as positive controls.

Four-micrometer-thick sections from paraffin-embedded tissues were cut and placed into 1.5-mL Eppendorf tubes. To avoid cross-contamination of samples, the microtome blade was cleaned with xylene between each block.20 DNA was extracted using EDTA-SDS/proteinase K treatment followed by phenol-chloroform, as previously described.8 DNA was resuspended with 50 μL of diethyl pyrocarbonate–treated and autoclaved pyrogen- and ribonuclease-free water, and 10 μL of extracted DNA was added to polymerase chain reaction (PCR) mixtures. To avoid false-negative results, nested PCR of the human androgen-receptor gene was performed in all cases as a positive control of DNA extraction.21 

To identify the presence of PV sequences, the DNA was subjected to multiplex nested PCR in order to amplify the large T antigen (LT) regions using the following primers: (1) PM1+ and PM1− as outer primers; and (2) PM2− (common to all PVs), JC+, BK+, and SV40+ as inner primers to distinguish different members of the Polyomavirus genus22 (Table 2). The amplification was performed on DNA extracted in a total volume of 50 μL with BioTaq DNA polymerase (Bioline, London, England) in the presence of 1:10 Bioline NH₄ buffer, 2 mM MgCl₂ (1 mM for the inner PCR), 10 pmol/μL of each primer (Roche Diagnostics, Milan, Italy), and 0.2 mM of dNTP, using a Progene Techno PCR System (Duotech, Milan, Italy). The samples were amplified by denaturation at 95°C for 5 minutes, followed by 40 cycles of denaturation at 95°C for 40 seconds, annealing at 61°C (55°C for the inner PCR) for 40 seconds, and extension at 72°C for 40 seconds. The cycles were terminated by means of a final extension at 72°C for 5 minutes. Diethyl pyrocarbonate–treated and ribonuclease-free water (Biotecx Laboratories, Houston, Tex) was used as negative control; the positive controls were DNA extracted from the brain tissue of subject 5 with PML (for JCV), the renal tissue of a subject with histologically proven BKV nephropathy (for BKV), and SVG cell lines (for SV40).

Histologic examinations of the kidney and urinary tract samples (see Table 1) revealed ischemic changes and nephroangiosclerosis in 8 cases (cases 2 through 5 and 7 through 10), acute tubular necrosis and kidney localization of Niemann-Pick disease in 1 (case 1), and renal bacterial abscess in 1 case (case 6). No changes were found in the renal pelvis, ureter, and urinary bladder samples. No morphologic equivalents of PV infection were identified in the specimens stained with hematoxylin and eosin.

Immunohistochemistry using antibodies against capsid proteins and LT antigens of PVs was always negative, as was in situ hybridization with probes recognizing JCV and BKV genomes.

DNA was obtained from all the examined samples, regardless of the time between death and autopsy.

On the whole, PV DNA was found in 43 (25.3%) of 170 samples from all patients. The frequency of PV detection in kidney and urinary tract segments is shown in Figure 1: urinary bladder (10/30 samples, 33%) and renal pelvis (6/20, 30%) were the main sites of PV DNA detection, whereas ureter and kidney tissue samples disclosed presence of PV DNA in 10 (25%) of 40 and 17 (21%) of 80 samples, respectively. Table 3 shows viral distribution and genotypes within kidney and urinary tract for each patient. Interestingly, PV DNA was detected in all of the subjects, with the number of positive samples ranging from 2 (cases 1 and 4) to 7 (case 6). As shown in Table 3, PV DNA was detected in only 1 case (case 3) in kidney and all segments of urinary tract; interestingly, PV was not detected more frequently in the patient with AIDS (case 5).

As shown in Figure 2, JCV was the genotype most frequently detected overall (23/43 positive samples, 53.5%) and within kidney (8/17 positive samples, 47%), renal pelvis (5/6 positive samples, 83%), and ureter (7/10 positive samples, 70%), whereas BKV was found in 14 samples (32.5%) and was the most frequent genotype found in the urinary bladder (6/10 positive samples, 60%). Coinfection by BKV-JCV was found on the whole in 6 samples (14%).

On the basis of the results of molecular studies of autopsy samples, the kidney is considered to be the main site of both JCV and BKV latency6,7; however, precise data concerning the distribution of PV infection in the kidney and the entire urinary tract are still lacking. Further evidence of the uropoietic system as the main site of PV latency is supplied by the frequency of hemorrhagic cystitis due to PV infection in patients with bone marrow transplants23,24 and the frequency of ureteral stenosis and PVN in patients with kidney transplants.12,14,15,25 It is not yet clear why PVN and ureteral stenosis are caused only by BKV infection18,25 and hemorrhagic cystitis by JCV or BKV damage to the urothelial cells of the urinary bladder,16,17 but one possible reason is a difference in cellular tropism for JCV and BKV; that is, the viruses may latently infect different segments of the uropoietic apparatus.

Although we studied only a small number of cases, the kidneys and urinary tracts were extensively sampled, and all samples underwent DNA extraction and amplification of the PV LT regions. Furthermore, the positive samples also underwent immunohistochemical examination and in situ hybridization. Our autopsy subjects were consecutive and unselected, and except for patient 5, who was affected by AIDS, all the others had no specific problems with immunosuppression; therefore, the subjects could be considered as representative of the general population.

Our results indicate that the kidney and the urinary tract are important sites of PV latency, because PV DNA was detected at these sites in all of our cases (with 2–7 positive samples for each autopsy). The urinary tract (renal pelvis, ureter, and urinary bladder) was the main site of viral detection, whereas PV DNA was found more rarely in the kidney (21% of cases). The latter finding confirms the Southern blot studies of Chesters et al,7 which revealed PV DNA in about 20% of cases; as far as we know, there are no published data concerning the distribution of PV in the urinary tract.

It is worth emphasizing that, in all of our cases (including the subject with AIDS), PV was present in a latent state, as indicated by the negative immunohistochemistry results. This analysis was performed using 1 antibody that recognizes viral capsid proteins (VP1–3) expressed only after virus assembly and an anti-LT antigen, which, although produced early after virus entry into the target cell, becomes detectable only during viral replication.2 The negative in situ hybridization results may have resulted from the fact that the number of copies in the cells was below the detection limit. False-negative results were avoided by the use of positive controls, including brain samples from the subject with PML (for JCV) and a kidney sample affected by BKV nephropathy.

JC virus was the main genotype identified overall and in kidney, renal pelvis, and ureter, whereas BKV prevailed in urinary bladder tissue. Because PVN and ureteral stenosis are caused by BKV reactivation and hemorrhagic cystitis by JCV and BKV, these results, although obtained in a small number of the cases, seem to indicate that both PVs persist randomly in the kidney and urinary tract and do not definitely explain the reasons for the different pathologic changes induced by JCV and BKV in the kidney and urinary tract.

However, some hypotheses can be tentatively proposed.

First of all, the sites of viral latency may be different from those in which PVs cause clinically relevant viral diseases: Nickeleit et al18 have suggested that in cases of PVN, viral particles produced after the lysis of infected cells in the renal pelvis or ureter could enter the capillary vessels and infect the tubular cells of the kidney, or that PVs may follow an ascending route of infection from transitional cells to kidney tubular cells.

Second, the susceptibility of transitional cells of the urinary tract and the tubular cells of the kidney to BKV or JCV infection may be different. For example, in patients with kidney transplants, JCV could damage the transitional cells of the urinary bladder but not the tubular cells of the kidney, where the virus seems to persist only in a latent state and does not cause renal disease.26 

Third, type and severity of immune impairment could be responsible for different diseases induced by PVs in the kidney and urinary tract. Indeed, we have demonstrated in a recent paper21 that, unlike patients who have undergone renal transplantation, in AIDS subjects JCV is able to replicate actively into the kidney epithelial tubular cells.

Finally, PV diseases in the kidney and urinary tract may be caused by genomic rearrangements in the regulatory region of viruses that latently infect the kidney and urinary tract. Because this region controls viral replication and infectiousness,27 different conditions of immunosuppression or changes in the tissue microenvironment (such as those occurring during renal transplantation) represent a possible mechanism of genomic instability, as proposed by Agostini et al.28 On the basis of this last hypothesis, the critical mechanisms may be more closely related to molecular viral damage than to the topographic distribution of latent virus within the kidney and urinary tract.

Further studies of larger case series, sequence analyses of isolated viral strains, or cloning and investigations on the distribution of PV in histologically normal-appearing tissues of patients affected by active PV diseases are needed to verify these hypotheses.